The applicants propose to develop and validate new image analysis methods aimed at a more accurate, reproducible, and automated approach to assessment of regional left ventricular (LV) function and visualization of 3D cardiac motion from tagged MRI data of patients with coronary artery disease (CAD). The applicants have developed a number of methods for analysis of tagged MRI data which have been validated in phantoms and animal models of myocardial infarction (MI). They propose to continue development of these techniques which utilize all of the available stripe information, including tag intersections and linear tag lines, in automatically taking LV deformations and reconstructing dense displacements at all myocardial points, with the goal of routinely applying these techniques to patient data. The advantage of the developed methods is that since displacement vectors will be available at all myocardial points, indices of LV function will also be available everywhere in the myocardium. These indices can be summed over local myocardial regions resulting in segmental function scores. In human studies, the developed methods will be applied to images acquired from normal volunteers, patients with pharmacologic stress-induced myocardial ischemia, patients with old, healed MI, and patients with ischemic dilated cardiomyopathy. In each case, segmental wall motion as assessed by the algorithms will be compared and correlated with validated clinical techniques such as 2D echocardiography, cine-MRI, and Gadolinium (Gd) contrast MRI. Thus, the specific aims are: (a) To measure statistical distribution (mean and standard deviation) of segmental function scores from 3D + t (short-axis and long-axis) tagged MRI at rest and under pharmacologic (dobutamine) stress in normal controls. (b) To measure the function scores as determined from 2D + t (short-axis) tagged MRI during pharmacologic stress and classified into normal, hypokinetic, or akinetic classes in patients with stress-induced ischemia. These labels will then be statistically correlated to labels assigned to the same segments by 2D echocardiography and cine-NIRI. (c) To measure segmental function scores as determined from 3D + t (short-axis and long-axis) tagged NIRI at rest and classified into normal, hypokinetic, akinetic, or dyskinetic classes in patients with an old, healed MI. The labels will be statistically compared to non-nal or akinetic labels assigned to the same segment from 2D echocardiography and cine-MRI, and with Gd contrast MRI. (d) To measure the segmental function scores from 3D + t (short-axis and long-axis) tagged MRI at rest and classified into normal, hypokinetic, akinetic, or dyskinetic classes in patients with ischemic, dilated cardiomyopathy. The labels will be statistically compared to labels assigned to the same segment from 2D echocardiography and cine-MRI.